I asked Claude, ChatGPT, and Gemini to fix the same bug, and only one understood it
… I put the "intuitive" intelligence of the models to test with 3 bugs Let the debugging games begin To test Sonnet 4.6, ChatGPT 5.5 and Gemini 3.1 , I sabotaged the physics of my Pygame project, "Captain Hat" with three logical bugs that break the platformer game's movement. …